Get Out of the Pool - A Look at Swiss-System Seeding

Note: This post was wtitten by Skylar Shibayama.

As exhilarating and well-run as Spikeball tournaments are, a glaring inefficiency remains in their format: pool play.

Pool play in its current form is lengthy, tedious, and does not fulfill its purpose of accurately ranking teams for bracket play. (Many teams recognize this and - smartly - conserve effort for the “real” part of the tournament.) By modifying the format of pool play we can see more of the energy and exciting matchups currently found only in the bracket stage. I propose that Spikeball use a “Swiss-system tournament” format to seed teams in place of pool play from here on out.

Mechanics and Background

The hallmark of the Swiss system is that for each round of play, the top-ranked team faces the #2 seed, the #3 faces the #4, #5 faces #6, and so on. After each round is completed, the entire field is then re-seeded for the next round (where again, #1 plays #2, #3 plays #4, etc.) Practically, this means that in the Swiss-style format, teams will (almost) exclusively face teams of equal record. This has numerous competitive and efficiency-related benefits that I will highlight later.

One might visualize this by imagining teams being continuously sorted into “pools” of same-record teams and each team plays another from their pool each round:

If there an odd number of teams, one random team should sit out each round (and never the same team twice.)

Seeds are determined using the following criteria:

1. W-L Record (Win%)
2. If tied, Strength of Schedule (Total # of Wins by Opponents - Total # of Losses by Opponents)
3. If tied, Total Point Differential (Points scored – Points allowed)
4. If tied, Random Tiebreak

In theory, Swiss-system play could continue for any number of rounds, but I suggest stopping once there is a single undefeated team (which will then be the #1 seed for the bracket.) That will be sufficiently accurate and avoid lots of repeat matchups. As you can see, this requires the same or fewer games than pool play, depending on division size:

*In an extremely large division, it might make sense to split teams into two big pools and complete use the Swiss system within each pool, eventually merging all teams back together for bracket play. 

The Swiss system is common across a wide variety of competitive activities, including chess (first in Zurich in 1895, hence the name,) debate, bowling, squash, and even the Pokémon Trading Card Game. Typically, the system replaces bracket play entirely; whoever is on top after a predetermined number of rounds is the champion. However, this can fail to replicate the excitement of a traditional bracket structure’s true championship game (since in theory the Swiss-system #1 seed could get so far ahead in the standings that their place would be cemented before the final round.) Therefore, I recommend Spikeball keep bracket play and simply use the Swiss system in place of pool play.

Last year at Birch Hill, Ezra and Ben Dantowitz ran pool play in this style successfully - accurate seeds were produced more efficiently than standard pool play. However, there weren’t enough Spikeball sets present for all teams to play at once. Since each round must be completed before determining the following round’s matchup, this caused unnecessary delays.

Since then, I have used the Swiss system in place of pool play in four tournaments, the largest being Northwest Spikefest at 25 teams (single division.) In each case, the Swiss-system seeding rounds took only about an hour, which gave us the time to include double-elimination bracket play and play out all seeds. I received exclusively positive feedback about the format, for a wide variety of reasons that I will detail.

Advantages

• Better Match-ups
• The most obvious draw of the Swiss system is that games are designed to include teams of similar skill. Compared to traditional pool play, there are fewer boring blowouts, making the tournament more watchable and worthwhile for all teams. With this brings more match-ups between top teams – if 2017 Nationals had used the Swiss system, for example, there is a good chance viewers would have gotten to see the hyped Cisek/Showalter v. Bud n’ Boles match-up without the two needing to meet in the bracket. Most importantly to the growth of the sport, the Swiss system ensures that, at most, only one team wins no games. In the current pool play format, a 50-team division can churn out seven or so teams that go home win-less, save for a post-tourney pickup game. The Swiss system would ensure that six of those seven finish with a win or two in a real counting match. This is good for the sport.
• More Efficient
• Swiss-system play slashes into long-dragging tournament days significantly and should take 90 minutes at the most to complete, even for gigantic divisions. Moreover, seeds are calculated automatically after each round, eliminating the time it normally takes to tabulate pool play results and set bracket seeds.

• Play out more seeds
• Double-elimination brackets
• Start later/finish earlier (and not have to worry about lighting)
  • Not have to cap pool play games at 15 and largely cause seven of the top eight Power Ranked teams to land on the same side of the 2017 Nationals bracket
• Play longer games or series
• Longer-than-recommended Swiss play for ultra-accurate seeds

• Better Assessor of Talent
• Swiss-system sorting removes pool play’s heavy reliance on Point Differential to seed teams. The #1 seed should be the team that can beat the other top competition, not the team that can beat bad teams by the most points. (This is true for all levels: middling teams should be ranked by their success against other middling teams, not their ability to lose to Cisek/Showalter by 8 instead of 10.) Additionally, this will help break the vicious cycle of deteriorating pool play effort that I see creeping in. When teams realize that pool play doesn’t accurately seed teams, they conserve energy because there is little penalty for receiving a poor seed. This then causes pool play to become even more inaccurate, leading to more teams conserving energy, and so on. (Which is fine if we want to just pre-seed or randomly seed brackets and have a warm-up before, but if we’re going to schedule the games they might as well be as watchable as possible.)

•  Easier for Tournament Directors
• Currently, Tournament Directors must effectively pre-seed every team in order to attempt to spread talent across pools evenly. Also, as stated before, adding up records and scores post-pool play is tedious – particularly if there are unequally-sized pools and thus a different number of games played to compare. Swiss play eliminates those unnecessary burdens and might allow TDs to sit down and actually enjoy their lunch every once in a while!

•  More Fair
• No longer are teams at the whim of the strength of their pool. All teams start on the exact same footing.

•  Less Time Sitting Around
• Depending on the number of teams in a division, there can be a lot of pool play byes. Take a 28-team division for example: four pools of seven will lead to four byes per round – that’s 14 missed opportunities to spike! The Swiss system needs no byes. (If there are an odd number of teams, only one bye per round is required.)

Disadvantages

•  Requires a Laptop
• Most Tournament Directors already use one, so this shouldn’t be an issue.

•  Requires Enough Nets/Space
• There must be enough nets and space for all teams to play at once; otherwise there is too much time spent waiting around and normal pool play might be more worthwhile.

•  Better Matchups
• In contrast to pool play, Swiss-system sorting won’t guarantee that all teams will face a top team. From the perspective of a bad team wanting to compete against an elite team, this is a downside. (I’d wager to guess the elite team might not see it the same way.) Regardless, the aforementioned advantages of better match-ups greatly outweigh this drawback in the name of entertainment and growth of the sport.

Conclusion

By replacing pool play with a Swiss-system format, we can make Spikeball tournaments more competitive, efficient, accurate, simple, fair, and lively with effectively zero downside. I urge the SRA Board to implement it in sanctioned events as soon as possible.

Of course, in order to utilize the format we must have quality automatic software, which I have created in Excel and am attaching here: https://drive.google.com/file/d/1SHl0jESUdaMw04YUtZcYvEWhvFQtZFn_/view

Please let me know if you have any feedback, positive or negative, either here, on Facebook, and/or at skylar.shibayama@aya.yale.edu. Spike on!